#pandaboard IRC Log on pandaboard.org

Join the chat at pandaboard.org/irc

Index

IRC Log for 2010-11-04

Timestamps are in UTC.

  1. [00:06:26] * peabody124 (~peabody12@adsl-99-41-234-41.dsl.hstntx.sbcglobal.net) Quit (Quit: peabody124)
  2. [00:09:50] * dm8tbr (dm8tbr@gw.bfst.de) Quit (Read error: Operation timed out)
  3. [00:09:58] * dm8tbr (dm8tbr@gw.bfst.de) has joined #pandaboard
  4. [00:13:43] * peabody124 (~peabody12@adsl-99-41-234-41.dsl.hstntx.sbcglobal.net) has joined #pandaboard
  5. [00:14:13] <topfs2> how do I make sure gstreamer uses openmax?
  6. [00:14:19] <topfs2> is there any log or something I can check ?
  7. [00:31:15] * th1a (~hoffman@pool-74-103-251-11.prvdri.fios.verizon.net) Quit (Quit: Ex-Chat)
  8. [00:33:28] * peabody124 (~peabody12@adsl-99-41-234-41.dsl.hstntx.sbcglobal.net) Quit (Quit: peabody124)
  9. [00:41:25] * NightMonkey (~NightMonk@pdpc/supporter/professional/nightmonkey) Quit (Read error: Connection reset by peer)
  10. [00:45:10] * j_ack (~Rudi@p57A42DAF.dip.t-dialin.net) Quit (Quit: Verlassend)
  11. [00:54:25] * NightMonkey (~NightMonk@pdpc/supporter/professional/nightmonkey) has joined #pandaboard
  12. [01:02:48] <topfs2> have the panda under your nose when you play with it can really make you afraid :) it smells abit when it gets hot :)
  13. [01:03:06] <topfs2> How hot should the chip really get? I'm able to have my finger on it for a little while but not extremely long
  14. [01:03:39] <mru> can you fry an egg on it?
  15. [01:03:59] <mru> quail egg, of course
  16. [01:04:15] <topfs2> I really doubt it :)
  17. [01:04:25] <mru> good
  18. [01:05:12] <topfs2> we both know how wasteful xbmc can be :) no way to get into an idle state ;)
  19. [01:05:32] <topfs2> the sgx core seems to be alot faster on this one htough
  20. [01:06:13] <topfs2> on beagle c4 playing SD content with our yuv-rgb shaders was at 2ish fps, now its full speed :)
  21. [01:06:27] <topfs2> Although, its really unecessary to use those but was fun trying to stress it abit
  22. [01:09:00] <topfs2> heck even running 1080p bunny is faster than running SD on beagle c4 before gsoc :)
  23. [01:13:31] * mpoirier (~quassel@S0106002369de4dac.cg.shawcable.net) Quit (Ping timeout: 240 seconds)
  24. [01:34:43] * alancam (~a-campbel@nat/ti/x-ebreztgssqvojyhy) Quit (Ping timeout: 240 seconds)
  25. [01:36:28] * vikpan (~a0876558@nat/ti/x-sxlzgpldvxqqrtun) has joined #pandaboard
  26. [01:36:30] <topfs2> is it ok to write into the /sys/devices/omapdss/display0/edid ? Seems like its hard to get it to go beyond 480p without a monitor
  27. [01:38:46] <topfs2> looks like thats a no
  28. [01:39:07] <topfs2> custom edid doent' seem to help without a screen
  29. [01:44:24] * vikpan (~a0876558@nat/ti/x-sxlzgpldvxqqrtun) has left #pandaboard
  30. [01:44:26] <mru> mine's doing 1080p nicely
  31. [01:46:42] * NightMonkey (~NightMonk@pdpc/supporter/professional/nightmonkey) Quit (Quit: Body blow! Body blow!)
  32. [01:48:01] <makmanalp> topfs2: so, it works? :P
  33. [01:48:13] <topfs2> mru, I'm guessing you have a screen though ;)
  34. [01:48:19] <topfs2> *monitor
  35. [01:48:21] <mru> yes
  36. [01:48:28] <topfs2> makmanalp, define works
  37. [01:48:36] <topfs2> And I still have not a full clue :)
  38. [01:48:41] <topfs2> can't test it all atm
  39. [01:48:47] <topfs2> gonna try getting omx working now though
  40. [01:49:05] <makmanalp> topfs2: neat
  41. [01:49:18] <topfs2> Seems really hard to get it to do anything besides failsafe 480p without monitor
  42. [01:49:23] <makmanalp> by works i meant 1080p, since that was what you were testing last i checked
  43. [01:49:26] <makmanalp> ahh
  44. [01:50:13] <topfs2> no thats not working (for me)
  45. [01:50:19] <topfs2> pretty sure it would work perfectly if I had a monitor :)
  46. [01:51:14] <mru> you could get one...
  47. [01:52:14] <topfs2> I have one at home, not sure I can afford one now for just a few months
  48. [01:52:25] <topfs2> but might just have to get oen
  49. [02:29:14] * emeb (~ericb@ip72-223-86-148.ph.ph.cox.net) has left #pandaboard
  50. [02:39:30] * FredFred (~FredFred@66-17-110-207-swiftelNET.brookings.net) has joined #pandaboard
  51. [02:54:18] * pcacjr (~pcacjr@187.78.112.52) Quit (Ping timeout: 245 seconds)
  52. [02:56:50] * pupnik (~puphome@p54B29D8E.dip.t-dialin.net) has joined #pandaboard
  53. [03:00:11] * pupnik_ (~puphome@p54867C44.dip.t-dialin.net) Quit (Ping timeout: 250 seconds)
  54. [03:00:13] * RekKumar (~a0322918@nat/ti/x-swekncdllpspsacr) has joined #pandaboard
  55. [03:14:23] <topfs2> ugh, looks like the omapdss just spews errors on omx without the screen (or I have done something bad)
  56. [03:14:36] <topfs2> seems like I need to get a screen, thankfully I've found a cheap 15" :)
  57. [03:32:29] * jonmasters (~jcm@dallas.jonmasters.org) has joined #pandaboard
  58. [03:32:49] <jonmasters> Good evening
  59. [03:37:38] * CarsonLynn (~carson@cpe-065-190-191-006.nc.res.rr.com) Quit (Ping timeout: 252 seconds)
  60. [03:45:46] * ajray (~ajray@short.csc.ncsu.edu) has joined #pandaboard
  61. [03:46:15] <ajray> quick question: Can the Pandaboard handle Ultra SDHC rates? like 15MBps writes?
  62. [03:46:45] <ajray> it says 'support for high-speed sdhc cards'
  63. [03:52:04] <ajray> i think the better answer is at those speeds I just need to get a USB hard drive.
  64. [04:14:50] * radhermit (~radhermit@gentoo/developer/radhermit) Quit (Read error: Operation timed out)
  65. [04:16:36] * radhermit (~radhermit@gentoo/developer/radhermit) has joined #pandaboard
  66. [04:42:19] * jprvita|afk (~jprvita@201.82.29.28) Quit (Ping timeout: 240 seconds)
  67. [05:09:32] * RekKumar1 (~a0322918@nat/ti/x-fahsfbjrjamdqziw) has joined #pandaboard
  68. [05:09:33] * RekKumar (~a0322918@nat/ti/x-swekncdllpspsacr) Quit (Remote host closed the connection)
  69. [05:10:02] * opm (~mike@col06-1-78-231-81-223.fbx.proxad.net) has joined #pandaboard
  70. [05:16:18] * FredFred (~FredFred@66-17-110-207-swiftelNET.brookings.net) Quit (Quit: FredFred)
  71. [05:25:31] * hrw|gone (~hrw@chello089073108238.chello.pl) Quit (Ping timeout: 240 seconds)
  72. [05:26:29] * hrw|gone (~hrw@chello089073108238.chello.pl) has joined #pandaboard
  73. [05:38:23] * RekKumar1 (~a0322918@nat/ti/x-fahsfbjrjamdqziw) Quit (Quit: Leaving.)
  74. [05:41:22] * radhermit (~radhermit@gentoo/developer/radhermit) Quit (Read error: Operation timed out)
  75. [05:43:22] * radhermit (~radhermit@gentoo/developer/radhermit) has joined #pandaboard
  76. [05:45:10] * thaytan (~jan@ppp59-167-167-201.static.internode.on.net) Quit (Ping timeout: 272 seconds)
  77. [05:57:36] * thaytan (~jan@ppp59-167-167-201.static.internode.on.net) has joined #pandaboard
  78. [06:48:06] * topfs2 (~topfs2@xbmc/staff/topfs2) Quit (Quit: Leaving)
  79. [07:05:36] * merbanan (~banan@c-83-233-242-220.cust.bredband2.com) Quit (Ping timeout: 272 seconds)
  80. [07:57:26] * lanthan_afh (~ze@p50992b91.dip0.t-ipconnect.de) has joined #pandaboard
  81. [08:08:16] * merbanan (~banan@c-94-255-221-206.cust.bredband2.com) has joined #pandaboard
  82. [08:08:59] * elesueur (~elesueur@2402:1800:4000:1:226:b0ff:feea:2790) Quit (Read error: Operation timed out)
  83. [08:21:44] * ndec (~ndec@nat/ti/x-weqtxkgedxysdjmz) has joined #pandaboard
  84. [08:27:59] * elesueur (~elesueur@120.20.31.159) has joined #pandaboard
  85. [08:37:09] * hrw|gone is now known as hrw
  86. [08:38:01] <hrw> moin
  87. [08:39:52] <eFfeM_work> hi hrw
  88. [08:46:16] * Ford_Prefect (~arun@gentoo/developer/ford-prefect) has joined #pandaboard
  89. [08:46:34] * Ford_Prefect (~arun@gentoo/developer/ford-prefect) has left #pandaboard
  90. [08:49:08] * Rhinebeck1 (~krishna@nat/ti/x-itprxvpguwcnevqs) Quit (Ping timeout: 240 seconds)
  91. [08:57:04] * RobotGuy (~n7pkt@c-24-21-60-36.hsd1.or.comcast.net) has joined #pandaboard
  92. [08:58:24] * fgu (~fgu___@cps2.dmz-eu.st.com) has joined #pandaboard
  93. [09:01:52] * opm (~mike@col06-1-78-231-81-223.fbx.proxad.net) Quit (Quit: Leaving)
  94. [09:20:14] * ahmedammar (~b33fc0d3@unaffiliated/b33fc0d3) has joined #pandaboard
  95. [09:21:55] * lrg (~lrg@slimlogic.co.uk) has joined #pandaboard
  96. [09:57:24] * florian_kc (~fuchs@Maemo/community/contributor/florian) has joined #pandaboard
  97. [09:58:19] * florian_kc is now known as florian
  98. [09:58:25] * ndec (~ndec@nat/ti/x-weqtxkgedxysdjmz) Quit (Remote host closed the connection)
  99. [10:04:51] * ndec (~ndec@nat/ti/x-vfyltxnuropxftev) has joined #pandaboard
  100. [10:08:19] * Monster (~carson@cpe-065-190-191-006.nc.res.rr.com) Quit (Ping timeout: 240 seconds)
  101. [10:19:33] * Monster (~carson@cpe-065-190-191-006.nc.res.rr.com) has joined #pandaboard
  102. [10:25:47] * ceyusa (~vjaquez@155.99.117.91.static.mundo-r.com) has joined #pandaboard
  103. [10:38:19] * hrw (~hrw@chello089073108238.chello.pl) Quit ()
  104. [10:38:47] * hrw (~hrw@chello089073108238.chello.pl) has joined #pandaboard
  105. [10:54:37] * kapejod (~kapejod@p57BB20C4.dip.t-dialin.net) has joined #pandaboard
  106. [10:58:14] * ceyusa (~vjaquez@155.99.117.91.static.mundo-r.com) Quit (Ping timeout: 264 seconds)
  107. [10:59:33] * ssvb (~ssvb@a88-114-220-213.elisa-laajakaista.fi) Quit (Quit: Leaving)
  108. [11:04:47] * ceyusa (~vjaquez@155.99.117.91.static.mundo-r.com) has joined #pandaboard
  109. [11:05:48] * ceyusa (~vjaquez@155.99.117.91.static.mundo-r.com) Quit (Client Quit)
  110. [11:12:09] <rsalveti> argh, got flooded by smsc95xx 1-1.1:1.0: usb0: kevent 2 may have been dropped
  111. [11:12:14] <rsalveti> while using usb as rootfs
  112. [11:17:10] * ceyusa (~vjaquez@155.99.117.91.static.mundo-r.com) has joined #pandaboard
  113. [11:37:58] * ian_brasil_ (~ian_brasi@189.2.128.130) has joined #pandaboard
  114. [11:40:08] * elesueur (~elesueur@120.20.31.159) Quit (Ping timeout: 245 seconds)
  115. [11:42:07] * jprvita|afk (~jprvita@187.106.5.201) has joined #pandaboard
  116. [11:42:37] * symm- (HydraIRC@178.168.29.31) has joined #pandaboard
  117. [11:46:10] * ian_brasil_ (~ian_brasi@189.2.128.130) Quit (Ping timeout: 272 seconds)
  118. [11:46:32] * ian_brasil_ (~ian_brasi@189.2.128.130) has joined #pandaboard
  119. [11:48:35] * ndec1 (~ndec@nat/ti/x-jwxujobdnkkhvujh) has joined #pandaboard
  120. [11:48:39] * ndec (~ndec@nat/ti/x-vfyltxnuropxftev) Quit (Quit: Leaving.)
  121. [12:12:11] * pcacjr (~pcacjr@187.112.11.189) has joined #pandaboard
  122. [12:12:13] * jprvita|afk is now known as jprvita
  123. [12:16:17] * simplon (8fcd7458@gateway/web/freenode/ip.143.205.116.88) has joined #pandaboard
  124. [12:16:53] <simplon> hi all! just a quick question: has anybody already received a pandaboard from digikey?
  125. [12:17:03] <av500> no
  126. [12:18:00] <simplon> do you know when they will start shipping?
  127. [12:18:28] <av500> no
  128. [12:20:16] * Monster (~carson@cpe-065-190-191-006.nc.res.rr.com) Quit (Ping timeout: 240 seconds)
  129. [12:27:57] * Monster (~carson@cpe-065-190-191-006.nc.res.rr.com) has joined #pandaboard
  130. [12:33:55] * gt11 (~geethat@12.157.84.42) has joined #pandaboard
  131. [12:36:58] * Monster (~carson@cpe-065-190-191-006.nc.res.rr.com) Quit (Ping timeout: 255 seconds)
  132. [12:38:12] * Monster (~carson@cpe-065-190-191-006.nc.res.rr.com) has joined #pandaboard
  133. [12:50:10] * symm- (HydraIRC@178.168.29.31) Quit (Ping timeout: 265 seconds)
  134. [12:54:16] * mpoirier (~quassel@S0106002369de4dac.cg.shawcable.net) has joined #pandaboard
  135. [12:57:33] * ogra_ac (~ogra@12.157.84.42) has joined #pandaboard
  136. [13:03:14] * fandeli (~fandeli@nat/ti/x-gfrskdwvlzhmhjpx) has joined #pandaboard
  137. [13:08:16] * robclark (~robclark@nat/ti/x-uxaehrynwjiuqvme) has joined #pandaboard
  138. [13:08:16] * ChanServ sets mode +o robclark
  139. [13:16:40] <sakoman> has anyone worked out which patches to mainline are required to get functioning mmc?
  140. [13:18:59] <mru> u-boot? kernel?
  141. [13:20:06] <sakoman> mru: kernel :-)
  142. [13:20:09] * topfs2 (~topfs2@xbmc/staff/topfs2) has joined #pandaboard
  143. [13:21:17] <mru> mainline as in torvalds?
  144. [13:21:38] <sakoman> mru: heh, I would be happy with either torvalds or l-o
  145. [13:22:02] * gt12 (~geethat@12.157.84.42) has joined #pandaboard
  146. [13:22:39] <sakoman> neither work at the moment
  147. [13:23:29] * gt11 (~geethat@12.157.84.42) Quit (Ping timeout: 255 seconds)
  148. [13:26:35] * jpinones-work (~x0066741@nat/ti/x-ophmoxyusflslmyv) has joined #pandaboard
  149. [13:26:50] <jpinones-work> hi
  150. [13:26:52] <jpinones-work> morning
  151. [13:27:54] * awayfar (~Adium@nat/ti/x-qaacdykfzflqdwws) has joined #pandaboard
  152. [13:29:20] * ogra_ac (~ogra@12.157.84.42) Quit (Ping timeout: 255 seconds)
  153. [13:37:16] <sakoman> gm jpinones-work
  154. [13:40:23] * ogra_ac (~ogra@12.157.84.42) has joined #pandaboard
  155. [13:42:06] * gt11 (~geethat@12.157.84.42) has joined #pandaboard
  156. [13:42:09] * gt12 (~geethat@12.157.84.42) Quit (Read error: Connection reset by peer)
  157. [13:43:26] * ogra_ac (~ogra@12.157.84.42) Quit (Read error: Connection reset by peer)
  158. [13:43:35] * ogra_ac_ (~ogra@12.157.84.42) has joined #pandaboard
  159. [13:48:23] * FredFred (~FredFred@66-17-110-207-swiftelNET.brookings.net) has joined #pandaboard
  160. [13:53:09] * gt12 (~geethat@12.157.84.42) has joined #pandaboard
  161. [13:53:18] * gt11 (~geethat@12.157.84.42) Quit (Quit: Leaving.)
  162. [13:56:47] * ogra_ac_ (~ogra@12.157.84.42) Quit (Ping timeout: 255 seconds)
  163. [14:00:34] * robclark (~robclark@nat/ti/x-uxaehrynwjiuqvme) Quit (Remote host closed the connection)
  164. [14:04:52] * BThompson (~a0193480@nat/ti/x-rtmcvupggcneyvqa) has joined #pandaboard
  165. [14:06:45] * simplon (8fcd7458@gateway/web/freenode/ip.143.205.116.88) Quit (Ping timeout: 265 seconds)
  166. [14:06:53] * opm (~mike@col06-1-78-231-81-223.fbx.proxad.net) has joined #pandaboard
  167. [14:09:36] * NishanthMenon (~nmenon@12.157.84.42) has joined #pandaboard
  168. [14:10:08] * jayabharath (~jayabhara@nat/ti/x-vqjhzuyrotfhaajo) has joined #pandaboard
  169. [14:17:44] * nhg (~a0864305@nat/ti/x-etfnlmjjzpgpfsgx) has joined #pandaboard
  170. [14:19:43] * pcarrier (~pcarrier@redhat/prc) has joined #pandaboard
  171. [14:19:48] <pcarrier> hello community :)
  172. [14:20:20] <pcarrier> can one achieve 1080p video playback currently?
  173. [14:20:35] <mru> what framerate?
  174. [14:20:51] <mru> and what codec?
  175. [14:21:23] <pcarrier> mru, 30fps, h264?
  176. [14:21:39] <mru> TI claims the iva can do it
  177. [14:21:43] <mru> I have yet to see it myself
  178. [14:21:57] <pcarrier> mru, so that's a no :)
  179. [14:22:13] <mru> I haven't even seen the code that's supposed to do it
  180. [14:22:54] <mru> I've asked for it several times but they keep pointing me at an ubuntu image
  181. [14:23:05] <mru> I'm sure you see the flaw in that
  182. [14:24:04] <topfs2> and I tried to start one but omapdss just complained and complained, probably due to no screen attached :)
  183. [14:24:30] <mru> topfs2: just go buy a monitor
  184. [14:25:02] <topfs2> will probably do that this weekend
  185. [14:25:12] <topfs2> still, would be nice if one weren't forced to
  186. [14:25:31] <mru> it's rather hard to watch video without one...
  187. [14:25:41] <av500> topfs2: switch omapdss to use a lcd screen instead of a monitor
  188. [14:25:46] <av500> lcd screen can talk back
  189. [14:25:48] <av500> cant
  190. [14:25:50] <av500> can't
  191. [14:26:06] <topfs2> av500, so that would probably mean xorg would use my modeline?
  192. [14:26:40] <topfs2> does sound like a smart suggestion, will try it
  193. [14:29:01] * ogra (~ogra@ubuntu/member/ogra) has joined #pandaboard
  194. [14:29:21] <av500> topfs2: i guess dss complains becasue hdmi detects no monitor
  195. [14:30:05] <topfs2> yeah, and it reverts to default res (640x480)
  196. [14:30:10] <topfs2> despite me having modeline in xorg.conf
  197. [14:33:53] <topfs2> So if I switch to omapdss.mode"dvi:1920x1080MR-24@60" that might do the trick?
  198. [14:34:03] <topfs2> Or you meant lcd not dvi?
  199. [14:34:21] <topfs2> omapdss.def_disp=lcd
  200. [14:35:38] <av500> try
  201. [14:36:42] <topfs2> on it
  202. [14:36:48] <topfs2> just wanted to make sure I understood you :)
  203. [14:41:13] * RekKumar (~a0322918@nat/ti/x-hawpcddcmjaaokwv) has joined #pandaboard
  204. [14:44:48] * pupnik (~puphome@p54B29D8E.dip.t-dialin.net) Quit (Quit: leaving)
  205. [14:45:13] <topfs2> dang
  206. [14:45:42] <av500> no, clang
  207. [14:45:49] <av500> easy to miss :)
  208. [14:46:15] <topfs2> no it didn't help:(
  209. [14:48:12] <jayabharath> pcarrier: yes most certainly can achieve 1080p playback with H264 or MPEG4 codec
  210. [14:48:25] <jayabharath> I am just demoing in now at LPC on a pandaboard
  211. [14:48:39] <jayabharath> with Ubuntu 10.10 stock image + TI addons loaded
  212. [14:48:52] <av500> jayabharath: that gave only dss underruns for me
  213. [14:48:57] <av500> even for SD :(
  214. [14:49:23] <jayabharath> I am able to playback a H264 clicp 24fps at 10Mb .. the theory says you can do 30fps
  215. [14:49:40] <topfs2> can I read the omapdss log somewhere?
  216. [14:49:48] <topfs2> to have some more clues?
  217. [14:49:48] <av500> enable omapdss debug
  218. [14:49:57] <topfs2> and it will print?
  219. [14:50:00] <av500> enable in kernel and pass on cmd line
  220. [14:50:06] <topfs2> in serial?
  221. [14:50:09] <av500> omapdss.debug=1 or so
  222. [14:50:11] <topfs2> will try
  223. [14:50:11] <jayabharath> av500: there is atleast 1 kernel paltch in the ubuntu kernel which makes the DSS underruns better.. also on ES2.1 there saying is that it's not as bad.. (but I dont know why)
  224. [14:50:24] <av500> jayabharath: why should it underrun at all?
  225. [14:50:31] <av500> even at 200mhz, with 2 sdram channels?
  226. [14:51:25] <av500> jayabharath: in my casem underun meant no image at all, that was using one of the funny ubuntu video players supllied in the image
  227. [14:52:18] <jayabharath> I havent digged into the issues - rob known a bit better - why' it occurs. But, what i hear is that the DSS internal buffers (fixed size) and not enough
  228. [14:52:35] <jayabharath> I just use the totem player in ubuntu
  229. [14:52:56] <av500> I will try again as soon as I can steal my PB away from my coworker
  230. [14:53:57] <mru> av500: come on... I know you're bigger
  231. [14:54:06] <mru> I doubt he'd put up a fight
  232. [14:58:11] * ogra (~ogra@ubuntu/member/ogra) Quit (Remote host closed the connection)
  233. [14:58:45] <jayabharath> av500: next time we ship you a assault rifle to protect your board ;)
  234. [14:59:29] <av500> please
  235. [14:59:34] <av500> should be no issue to get in TX
  236. [14:59:55] <jannau> jayabharath: that will only cause trouble at the german customs
  237. [15:00:10] * ogra_ac (~ogra@12.157.84.42) has joined #pandaboard
  238. [15:00:21] <mru> shotgun is probably better in those close-quarters office battles
  239. [15:00:29] * pupnik (~puphome@p54B29D8E.dip.t-dialin.net) has joined #pandaboard
  240. [15:00:34] * opm (~mike@col06-1-78-231-81-223.fbx.proxad.net) Quit (Quit: Leaving)
  241. [15:00:55] <av500> we already had a small skirmish about the office tree today
  242. [15:06:05] * ogra_ac (~ogra@12.157.84.42) Quit (Ping timeout: 255 seconds)
  243. [15:06:56] <pupnik> i have a nice app idea for the DSP
  244. [15:07:33] <av500> process digital signals?
  245. [15:09:32] <pupnik> was just looking at electric pianos, for sequencing music to midi. pricey.
  246. [15:11:03] <pupnik> but for someone who already has a piano....
  247. [15:13:48] <jayabharath> pupnik: I would not venture to develop for the DSP (mini 64x) on OMAP4... OMAP3 (real 64x) DSP is much better
  248. [15:14:40] <jayabharath> The other problem is the DSP bridge stuff is not really supported anymore - I was talking to the syslink guys.. they dont want to support it ;)
  249. [15:15:02] <av500> jayabharath: err, there will be some kind of access to the dsp? or none at all?
  250. [15:15:15] <jayabharath> JTAG ;)
  251. [15:15:41] <av500> just hook a few gpios :)
  252. [15:15:49] <av500> and do it from the a9...
  253. [15:15:58] <jayabharath> The SW stack from TI does not have DSP bridge or anything running on the DSP...
  254. [15:16:09] <av500> so what runs on the dsp?
  255. [15:16:14] <av500> idle loop?
  256. [15:16:28] <pupnik> i have only met one person who has managed to learn TI's 64x and make a FOSS app using it
  257. [15:16:39] <jayabharath> There is work to leverage the mini-C64x for Gestures libraries, Natural UI accelerations etc
  258. [15:16:54] <av500> as in closed api from the A9
  259. [15:17:01] <pupnik> very nice jayabharath - that interactivity stuff would be great to pull off the main cpu
  260. [15:17:10] <jayabharath> Yes
  261. [15:17:13] * j_ack (~Rudi@p57A42167.dip.t-dialin.net) has joined #pandaboard
  262. [15:17:21] <av500> jayabharath: gestures, no where have I heard that
  263. [15:17:24] <av500> now
  264. [15:17:46] <jayabharath> yeap it a work that is in progress...
  265. [15:18:12] <jayabharath> details are not in public yet
  266. [15:18:35] <av500> make sure there is a cam to attach at least :)
  267. [15:18:46] <av500> gestures from still images is a bit lame :)
  268. [15:19:11] <jayabharath> Most of this development is on a Blaze platform.. which has 3 cameras
  269. [15:19:31] <jayabharath> Also, you can add ultrasonic mics etc that usually are used for this stuff
  270. [15:19:57] <pupnik> i think the application developers need a standard way to put monitor/wakeup code on a small cpu that consumes almost no energy - keep the main core and memory in sleep
  271. [15:20:38] <jayabharath> BTW, there is a linux port to c6x DSP... it might be interesting to run linux on that and use some IPC to talk to the linux on ARM
  272. [15:20:55] <av500> i fail to see the point of that
  273. [15:21:23] <topfs2> av500, where are the logs supposed to happen? in serial output?
  274. [15:21:30] <av500> printk
  275. [15:21:33] <av500> try dmesg
  276. [15:21:46] <pupnik> av500: because peripherals and radios consume so much more energy?
  277. [15:21:49] <av500> dmesg -n 8
  278. [15:22:00] <av500> pupnik: ?? I meant the linux on the dsp
  279. [15:22:21] <topfs2> now there we have something, thanks av500
  280. [15:24:18] <topfs2> [ 2.463531] omapdss HDMI: No edid set thus will be calling hdmi_read_edid
  281. [15:24:18] <topfs2> [ 2.473602] hdmi: Analog Display
  282. [15:24:18] <topfs2> [ 2.473602] hdmi: EDID Content 0
  283. [15:24:18] <topfs2> [ 2.473632] omapdss HDMI: fallback to VGA
  284. [15:24:37] <av500> yes
  285. [15:24:42] <mru> analog wtf?
  286. [15:24:43] <av500> vga is the hdmi fallback
  287. [15:24:59] <topfs2> and then it powers mode 40, which makes sense
  288. [15:25:11] <topfs2> Not sure if I can forcefully make it use mode 51
  289. [15:25:14] <av500> mru: yes, hdmi thinks an analog display is connected as it does not reply on the edid
  290. [15:25:27] <mru> but hdmi has no analog pins...
  291. [15:25:27] <av500> i guess...
  292. [15:25:37] <av500> mru: maybe the driver does not know :)
  293. [15:25:41] <mru> and analog displays often do have edid
  294. [15:25:54] <mru> it's in some variant of the vga spec
  295. [15:26:27] <topfs2> Seems like it doesn't do anything with the lcd
  296. [15:26:41] <av500> topfs2: could well be that this is not supported yet
  297. [15:26:49] <av500> i heard something like only hdmi fttb
  298. [15:27:04] <topfs2> oh, noticed that on pandaboard wiki it said only hdmi out yeah
  299. [15:27:11] <topfs2> i.e. dvi didn't work, so might be the same with hdmi
  300. [15:27:16] <topfs2> *lcd
  301. [15:27:25] <av500> yes, same dss output port
  302. [15:27:29] <av500> like on BB
  303. [15:27:31] <mru> I didn't get anything on the dvi output
  304. [15:27:39] <mru> but then I didn't try very hard
  305. [15:27:40] <topfs2> oh, that explains it then :)
  306. [15:27:42] <mru> since hdmi works fine
  307. [15:27:42] <av500> hdmi is the new thing for omap4 dss
  308. [15:27:44] * orbarron (~orb@nat/ti/x-qvhohroywtrzofss) has joined #pandaboard
  309. [15:27:44] * ChanServ sets mode +o orbarron
  310. [15:27:56] <av500> topfs2: canada has no hdmi/dvi monitors?
  311. [15:28:06] <topfs2> they do but I'm a poor student :)
  312. [15:28:10] <av500> us trade embargo?
  313. [15:28:15] <mru> av500: dvi is not allowed in the colonies
  314. [15:28:24] <av500> right
  315. [15:28:35] <av500> topfs2: go throw tea into the harbour
  316. [15:29:02] <topfs2> lol what?
  317. [15:29:58] <mru> boston tea party
  318. [15:30:10] <av500> and bing, no more a colony
  319. [15:30:15] <mru> didn't you pay attention during history class?
  320. [15:30:28] <av500> or pay somebody else to pay attention at least
  321. [15:30:31] <topfs2> not sure how thta would bring me a tv though :)
  322. [15:30:33] <mru> or have .se schools degraded that much since I attended?
  323. [15:30:38] <topfs2> cat manager0/name
  324. [15:30:38] <topfs2> lcd
  325. [15:30:38] <topfs2> looks like something happened
  326. [15:36:33] * emeb (~ericb@ip72-223-86-148.ph.ph.cox.net) has joined #pandaboard
  327. [15:38:33] <pcarrier> jayabharath, so how's the hardware video decompression handled?
  328. [15:38:41] <pcarrier> jayabharath, does it work in mplayer/xine/gstreamer?
  329. [15:39:26] <topfs2> openmax
  330. [15:39:58] <jayabharath> gstreamer-> openmax -> accelerators ...
  331. [15:40:01] * prpplague (~danders@nat/ti/x-mkltbwamvaerhalt) has joined #pandaboard
  332. [15:40:10] <jayabharath> prpplague: morning
  333. [15:40:15] <av500> jayabharath: so gst is the only SW that can use omx?
  334. [15:40:31] * CarsonLynn (~carson@nom22900b.nomadic.ncsu.edu) has joined #pandaboard
  335. [15:40:44] <prpplague> jayabharath: greetings
  336. [15:41:09] <jayabharath> not necessarily -- you can call into the OMX libs directly we have gst plugins that use the omx stuff .. as gst is more standard way to talk to a varienty of media apps
  337. [15:42:33] <mru> now that's the attitude I despise
  338. [15:44:38] <noleguy> Just an fyi but an updated Pandroid Froyo for 27.6.2 is available for download on http://code.google.com/p/pandroid/
  339. [15:44:47] * CarsonLynn (~carson@nom22900b.nomadic.ncsu.edu) Quit (Ping timeout: 255 seconds)
  340. [15:45:31] <av500> noleguy: thx
  341. [15:46:56] <noleguy> No GFX, Ducati, or WLAN drivers are bundled. They have to be downloaded, lic agreement accepted, and installed separately, etc....
  342. [15:47:24] * jpinones-work (~x0066741@nat/ti/x-ophmoxyusflslmyv) has left #pandaboard
  343. [15:47:34] <av500> noleguy: as long as it is explaind
  344. [15:47:39] * jpinones-work (~x0066741@nat/ti/x-ophmoxyusflslmyv) has joined #pandaboard
  345. [15:48:03] <noleguy> Lastly, these are some pre-built binaries for a generic setup, etc...
  346. [15:48:18] <noleguy> There is a README in the kernel download.
  347. [15:51:07] * Rhinebeck1 (~krishna@nat/ti/x-hlfzkdymogaztles) has joined #pandaboard
  348. [15:54:19] * robclark (~robclark@nat/ti/x-ywkstkqbznfdazvn) has joined #pandaboard
  349. [15:54:19] * ChanServ sets mode +o robclark
  350. [15:55:01] * nhg (~a0864305@nat/ti/x-etfnlmjjzpgpfsgx) has left #pandaboard
  351. [16:02:25] * ChanServ sets mode +o prpplague
  352. [16:03:48] <guerby> jayabharath, Fedex just delivered the GCC compile farm pandaboard, thanks!
  353. [16:03:51] <guerby> armin76, ^
  354. [16:03:57] <guerby> now I need a power supply
  355. [16:04:04] <jayabharath> great
  356. [16:04:15] <jayabharath> yeah the power supply thing is a bit of a pain
  357. [16:04:35] <jayabharath> perhaps in future we can try to ship a y cable after we verify most features of boards works over USB
  358. [16:04:39] <av500> use and TI EVM power supply
  359. [16:04:42] <av500> any
  360. [16:08:30] <armin76> guerby: are those thanks for me too? :D
  361. [16:08:52] <guerby> armin76, yep :)
  362. [16:08:55] <armin76> guerby: fyi i'm using a 2a power supply until it arrives the 4a one i bought on ebay
  363. [16:09:13] <guerby> ok I have a 5V 2.4A but the connector doesn't fit
  364. [16:09:23] <armin76> too bad
  365. [16:09:40] <guerby> armin76, is there a name for the connector ?
  366. [16:10:35] * ndec1 (~ndec@nat/ti/x-jwxujobdnkkhvujh) Quit (Ping timeout: 250 seconds)
  367. [16:10:35] <armin76> guerby: well, there are size specifications
  368. [16:10:47] <guerby> jayabharath, there's a little bit of "glue" like substance on the wifi ship, I assume the sticker there went away
  369. [16:11:20] <jayabharath> yes.. the sticker had to be removed when flowing through the assembly ovens
  370. [16:11:22] <prpplague> guerby: we have had issues with the sticker melting off during the assembly of the board
  371. [16:11:26] <jayabharath> so thats normal
  372. [16:11:43] <jayabharath> prpplague: the 10" display we have what's it resolution.. android does not report it
  373. [16:12:02] <prpplague> 1024x600
  374. [16:12:05] <armin76> guerby: http://search.digikey.com/scripts/DkSearch/dksus.dll?lang=en&site=US&WT.z_homepage_link=hp_go_button&KeyWords=T951-P5P-ND+&x=0&y=0 <- check the 'cord plug' specs
  375. [16:12:08] <guerby> jayabharath, ok thanks for the confirmation
  376. [16:12:28] <prpplague> jayabharath: pretty much the standard resolution for 10.1" netbooks
  377. [16:12:35] <guerby> armin76, Cord Plug 2.1mm I.D. x 5.5mm O.D. x 12mm Female
  378. [16:12:37] <guerby> ok noted
  379. [16:13:17] <guerby> http://pandaboard.org/content/resources/Accessories
  380. [16:13:21] <armin76> guerby: if it helps, i'm using a power supply from a dlink access point
  381. [16:13:41] <jayabharath> prpplague: ok. thanks
  382. [16:14:03] * Sowa (~yaaic@idmm.gprs.19522256-76.bih.net.ba) has joined #pandaboard
  383. [16:14:05] <armin76> guerby: if you want i can give you the link of the power supply i bought from ebay, from uk
  384. [16:14:19] <armin76> shipping from digikey was 30$ :/
  385. [16:14:35] * gt12 (~geethat@12.157.84.42) Quit (Quit: Leaving.)
  386. [16:15:04] <prpplague> jayabharath: did you get some questions about it?
  387. [16:15:10] <guerby> armin76, yep send it to me
  388. [16:15:17] <jayabharath> yes like.. where did you get the tape?
  389. [16:15:23] <guerby> armin76, jayabharath the connector should look like that right ? http://www.transistek.com/catalogue/connecteurs_adaptateurs/connecteurs_secteur/fiches_d_alimentation_cc/fiches_femelles/connecteur_d_alimentation_cc_femelle_2_1mm_x_5_5mm_150.php
  390. [16:15:31] <jayabharath> People want acrylic cases ;)
  391. [16:15:52] * gt11 (~geethat@12.157.84.42) has joined #pandaboard
  392. [16:16:05] <prpplague> jayabharath: hehe
  393. [16:16:22] <av500> cases are overrated
  394. [16:16:48] <jayabharath> av500: some software guys get scared when seeing a bare PCB.. they just want to hide it in a case ;)
  395. [16:17:00] <av500> well, teach 'em
  396. [16:17:04] * NightMonkey (~NightMonk@pdpc/supporter/professional/nightmonkey) has joined #pandaboard
  397. [16:17:12] <armin76> guerby: hrm...in fact the power supply needs to be a male connector, since the panda is female
  398. [16:17:14] <av500> make em wear wireless emc wriststraps too
  399. [16:17:23] <armin76> jayabharath: am i right?
  400. [16:17:26] <jayabharath> We cant teach a cat to give birth to a lion ;)
  401. [16:17:28] <guerby> armin76, stoopid me
  402. [16:17:47] * jayabharath looks
  403. [16:18:12] * fandeli (~fandeli@nat/ti/x-gfrskdwvlzhmhjpx) Quit (Quit: ChatZilla 0.9.86 [Firefox 3.6.12/20101027124711])
  404. [16:18:13] <armin76> guerby: check the pics on the digikey page
  405. [16:18:17] <guerby> http://www.transistek.com/catalogue/connecteurs_adaptateurs/connecteurs_secteur/fiches_d_alimentation_cc/fiches_males/fiche_d_alimentation_male_2_1_x_5_5_x_14mm_140.php
  406. [16:18:59] <armin76> i think that would fit, not sure
  407. [16:19:03] <guerby> armin76, ok there's a drawing in the spec sheet
  408. [16:19:20] <jayabharath> guerby: this one should fit...
  409. [16:19:27] <jayabharath> but not 100% sure
  410. [16:19:29] * ozk (~oz@nat/ti/x-zgnwusvubjwjnyba) has joined #pandaboard
  411. [16:19:34] <orbarron> woohoo... looks like pandaboard booted with Ubuntu 10.10 with dual usb cable -- no power supply needed -- not sure what else I can run but at least this is up
  412. [16:19:58] * Sowa (~yaaic@idmm.gprs.19522256-76.bih.net.ba) Quit (Quit: Yaaic - Yet another Android IRC client - http://www.yaaic.org)
  413. [16:20:34] <orbarron> mouse, keybaord, ethernet, and hdmi all up and running
  414. [16:20:58] <jayabharath> orbarron: how much power does the minimal fs consume?
  415. [16:21:28] * orbarron switches set up -- give me a min
  416. [16:21:35] <av500> a fs does does not consume power
  417. [16:21:39] <armin76> lol
  418. [16:21:50] <jayabharath> :D
  419. [16:22:03] * ogra_ac (~ogra@12.157.84.42) has joined #pandaboard
  420. [16:24:19] <pupnik> are there better sources of hires lcds than digikey?
  421. [16:24:36] <av500> I can offer you 1024x600 for $300
  422. [16:24:51] <prpplague> don't listen to av500! hehe
  423. [16:24:54] <av500> with capacitive touch
  424. [16:25:04] <pupnik> what size would that be?
  425. [16:25:09] <av500> 10"
  426. [16:25:13] <av500> 10.1"
  427. [16:25:14] * makmanalp (~legato@constant.inople.net) Quit (Ping timeout: 264 seconds)
  428. [16:25:30] <av500> you'd need to rip it out of an otherwise fine product though :)
  429. [16:26:16] <pupnik> maybe my cheapest bet is a cheap chinese android 1.6 tablet
  430. [16:26:22] <dm8tbr> pfff just wait for SDE, then you can use it almost like a beagle-board :)
  431. [16:26:38] * ogra_ac (~ogra@12.157.84.42) Quit (Ping timeout: 255 seconds)
  432. [16:27:12] <pupnik> the cheapest tablets don't even have 16-bit displays though - unusable really
  433. [16:27:45] <pupnik> have any of you desoldered and replaced that ethernet/usb tower yet?
  434. [16:28:04] <guerby> ok I found a power supply with the right plug
  435. [16:28:20] <pupnik> :) /me applauds good old guerby
  436. [16:28:48] <guerby> but not the right power spec 6V 800 mA
  437. [16:29:12] <orbarron> jayabharath: hmm so I am see ~500mA idle off of 5v source
  438. [16:30:50] <jayabharath> orbarron: ?
  439. [16:31:08] <av500> guerby: put a 1V zener and a large heatsink
  440. [16:36:26] * makmanalp (~legato@constant.inople.net) has joined #pandaboard
  441. [16:38:15] * ogra_ac (~ogra@12.157.84.42) has joined #pandaboard
  442. [16:46:49] * davidc__ (~davidc__@S0106000f662a10a4.vc.shawcable.net) Quit (Quit: davidc__)
  443. [16:47:47] * ogra_ac (~ogra@12.157.84.42) Quit (Ping timeout: 255 seconds)
  444. [16:48:31] <orbarron> jayabharath: yup... ~500mA is what I am seeing
  445. [16:49:07] <jayabharath> Ok. Thanks..
  446. [16:49:25] <jayabharath> how does it jump up when you scan & connect to a WLAN AP
  447. [16:49:51] <av500> like this: ___/\____
  448. [16:51:00] <jayabharath> ;)
  449. [16:51:10] * opm (~mike@col06-1-78-231-81-223.fbx.proxad.net) has joined #pandaboard
  450. [16:51:13] <mru> not like this: _____|_____
  451. [16:51:15] <mru> ?
  452. [16:52:33] <av500> i used a different time base
  453. [16:52:46] <av500> i have a faster asciiscope
  454. [16:53:00] <mru> a proper dirac pulse is independent of time base
  455. [16:53:30] <av500> yes, but men in black came and took away my ideal dirac generator
  456. [16:53:49] <av500> now I only have firefox reload button
  457. [16:54:01] <jpinones-work> F5?
  458. [17:01:38] * davidc__ (~davidc__@S0106000f662a10a4.vc.shawcable.net) has joined #pandaboard
  459. [17:12:29] * jayabharath (~jayabhara@nat/ti/x-vqjhzuyrotfhaajo) Quit (Remote host closed the connection)
  460. [17:16:10] * gt11 (~geethat@12.157.84.42) Quit (Quit: Leaving.)
  461. [17:20:46] * sum (~11root11@fiesta.cs.tu-berlin.de) Quit (Read error: Operation timed out)
  462. [17:22:29] * NishanthM (~nmenon@12.157.84.42) has joined #pandaboard
  463. [17:23:23] * NishanthMenon (~nmenon@12.157.84.42) Quit (Ping timeout: 250 seconds)
  464. [17:23:49] * lanthan_afh (~ze@p50992b91.dip0.t-ipconnect.de) Quit (Quit: Ex-Chat)
  465. [17:24:17] * jayabharath (~jayabhara@nat/ti/x-zfmfaszipmchljeq) has joined #pandaboard
  466. [17:26:53] * sum (~11root11@fiesta.cs.tu-berlin.de) has joined #pandaboard
  467. [17:27:19] * sum is now known as Guest63722
  468. [17:30:09] * gt11 (~geethat@12.157.84.42) has joined #pandaboard
  469. [17:33:39] * hrw is now known as hrw|gone
  470. [17:40:55] * florian (~fuchs@Maemo/community/contributor/florian) Quit (Quit: Client exiting)
  471. [17:47:54] * Vikas (~a0868693@nat/ti/x-qwdyjwkmbrlswjsd) has joined #pandaboard
  472. [17:50:22] <jayabharath> orbarron: ping
  473. [17:53:36] <mru> jayabharath: who's the person to talk to about getting the full omap4 trm?
  474. [17:57:07] * th1a (~hoffman@pool-74-103-251-11.prvdri.fios.verizon.net) has joined #pandaboard
  475. [17:59:47] * Vikas (~a0868693@nat/ti/x-qwdyjwkmbrlswjsd) Quit (Ping timeout: 250 seconds)
  476. [18:02:23] * cooloney (~roc@12.157.84.6) has joined #pandaboard
  477. [18:03:31] * ogra (~ogra@ubuntu/member/ogra) has joined #pandaboard
  478. [18:11:16] * ogra (~ogra@ubuntu/member/ogra) Quit (Ping timeout: 240 seconds)
  479. [18:12:10] <jayabharath> me ;)
  480. [18:14:10] * mru talks
  481. [18:15:42] * CarsonLynn (~carson@nom27910d.nomadic.ncsu.edu) has joined #pandaboard
  482. [18:16:51] * fandeli (~fandeli@nat/ti/x-nnwavbgucjhrhbjk) has joined #pandaboard
  483. [18:18:33] * ogra_ac (~ogra@12.157.84.42) has joined #pandaboard
  484. [18:19:48] * Monster (~carson@cpe-065-190-191-006.nc.res.rr.com) Quit (Ping timeout: 240 seconds)
  485. [18:20:30] * gxben (~gxben@plo67-1-88-177-29-32.fbx.proxad.net) has joined #pandaboard
  486. [18:20:54] * Monster (~carson@cpe-065-190-191-006.nc.res.rr.com) has joined #pandaboard
  487. [18:23:07] <FredFred> mru get a pandaboard
  488. [18:23:20] <FredFred> oopse...what's trm?
  489. [18:24:03] <jayabharath> trm= Techical Reference Manual http://tinyurl.com/omaptrm
  490. [18:24:33] <FredFred> tnx!
  491. [18:25:22] * Monster (~carson@cpe-065-190-191-006.nc.res.rr.com) Quit (Ping timeout: 252 seconds)
  492. [18:27:19] * VikasJoshi (~a0868693@12.157.84.42) has joined #pandaboard
  493. [18:27:33] * J_P (~neo@189.2.146.53) has joined #pandaboard
  494. [18:27:37] <J_P> hi all
  495. [18:28:19] <VikasJoshi> J_P Hi what's cooking
  496. [18:29:55] <FredFred> Anyone have any inside info on how many pandaboards are backordered at Digi-Key and/or how many they'll be getting in on the 21st?
  497. [18:30:32] <J_P> VikasJoshi: humm? what?
  498. [18:31:07] * Monster (~carson@cpe-065-190-191-006.nc.res.rr.com) has joined #pandaboard
  499. [18:32:08] <VikasJoshi> FredFred: I think you order yours and you should get yours on 21st... to my knowledge digikey will be shipping them in 3rd week of Nov (21st as you mention)
  500. [18:32:15] <VikasJoshi> J_P: are you
  501. [18:33:20] * pcacjr (~pcacjr@187.112.11.189) Quit (Ping timeout: 260 seconds)
  502. [18:34:01] <armin76> prpplague: is the stuff at gitorius going to get updated?
  503. [18:34:53] <J_P> VikasJoshi: I would like to development a board with video chips capture to capture images from analogue cameras. So I was thinking about chip BT878 that is cheap and good image. Questio is if is possible to use bt878 for that.. because I thinking to use V4l2 API with OMAP4.
  504. [18:36:35] <mru> J_P: bt878 has a pci interface, omap4 doesn't
  505. [18:37:05] <J_P> mru: maybe saa?
  506. [18:37:14] <mru> the philips chips?
  507. [18:37:23] <mru> anything you'll find in a pc is pci
  508. [18:37:29] <J_P> mru: yes.. or tell me a good chip..
  509. [18:37:38] <mru> there are for sure other chips available
  510. [18:37:43] <mru> but I don't know them
  511. [18:37:50] <VikasJoshi> J_P: bt878 as I know has pci
  512. [18:38:00] <mru> VikasJoshi: yes, that's the point
  513. [18:38:01] <VikasJoshi> J_P: • Supports NTSC/PAL/SECAM video decoding
  514. [18:38:01] <VikasJoshi> • Supports image resolutions up to 768x576 (full
  515. [18:38:01] <VikasJoshi> PAL resolution)
  516. [18:38:01] <VikasJoshi> • Supports complex clipping of video source
  517. [18:38:01] <VikasJoshi> • Zero wait state PCI burst writes
  518. [18:38:17] <mru> VikasJoshi: there's no way to connect a pci chip to the omap4
  519. [18:38:22] <mru> no sane way at least
  520. [18:38:38] <mru> you'd have to build a pci interface for it first
  521. [18:38:51] <mru> which you could put on gpmc or c2c or course
  522. [18:38:59] <mru> but that's just a lot of work
  523. [18:39:02] <jayabharath> armin76: it should (that's the plan) .. keep pushing prpplague to get it done fast..
  524. [18:39:03] <VikasJoshi> But omap4 doesnt but i am thinking of any other way .... are you hard on bt878?
  525. [18:39:07] <mru> better to get something that can feed the image input directly
  526. [18:39:17] <J_P> mru: what is easier way for that?
  527. [18:39:33] <J_P> mru: I use bt878 in pc..
  528. [18:39:40] <mru> PCs have pci bus
  529. [18:39:51] <armin76> jayabharath: well, tbh its not funny that there's no 'official tree' updated enough
  530. [18:40:15] <armin76> but for ubuntu there is...
  531. [18:41:00] <jayabharath> The git.omapzoom.org tree will usually be ahead of most trees... (except kernel.org -- which is not fully functional on panda)
  532. [18:41:33] <jayabharath> I believe ubuntu kernel will only move with ubuntu releases
  533. [18:41:37] <armin76> jayabharath: you mean the integration one?
  534. [18:41:44] <jayabharath> if you need a more modren kernel use the integration one
  535. [18:42:04] * ogra_ac (~ogra@12.157.84.42) Quit (Read error: Connection reset by peer)
  536. [18:42:07] <jayabharath> the gitorious is a rebased of the integration one + a bunch of patches + anything else anyone will contribu to it
  537. [18:42:34] <J_P> mru: what you think is easier way for that? I would like to do with OMAP4 like as this board: http://www.lex.com.tw/product/pdf/SI525AW.pdf
  538. [18:42:53] <prpplague> armin76: is there some aspect of the gitorious trees you are looking at?
  539. [18:43:24] <armin76> prpplague: the alsa thing f.ex
  540. [18:43:25] <jayabharath> prpplague: I guess he is pointing to the fact that there should be 'a' tree that is alwasy uptodate with latest and greatest
  541. [18:43:54] <jayabharath> today some changes in ubuntu kernel.. not in gitorious.. some in integration tree.. i.e., gradually we are diverging code bases.. not good
  542. [18:43:55] <armin76> and the uboot from there doesn't support 1gb of ram
  543. [18:43:58] <mru> that would indeed be great
  544. [18:44:46] <prpplague> jayabharath: i guess that goes to do we just mirror whats on ubuntu's git as well as the dev.omapzoom as welll?
  545. [18:44:52] * ogra_ac (~ogra@12.157.84.42) has joined #pandaboard
  546. [18:45:05] <prpplague> armin76: i am no aware of the "alsa thing"
  547. [18:45:12] <prpplague> armin76: can you point me to some info?
  548. [18:45:48] <armin76> prpplague: sure, gimme a sec
  549. [18:46:00] <jayabharath> prpplague: yeah... just to grab and collect all the valid patchsets into gitorious tree -- do that make sense?
  550. [18:46:11] <ogra_ac> its a sound driver layer for linux systems
  551. [18:46:22] <armin76> https://bugs.launchpad.net/ubuntu/+source/alsa-utils/+bug/637947
  552. [18:47:11] <prpplague> jayabharath: yea, but base it on which git tree? all of them?
  553. [18:47:15] <ogra_ac> prpplague, your panda breadbox is awesome btw ;)
  554. [18:47:27] * VikasJoshi (~a0868693@12.157.84.42) has left #pandaboard
  555. [18:47:28] * Monster (~carson@cpe-065-190-191-006.nc.res.rr.com) Quit (Ping timeout: 240 seconds)
  556. [18:47:33] <J_P> mru: did you see that link
  557. [18:47:33] <J_P> ?
  558. [18:48:12] <armin76> prpplague: i see stuff updated for alsa on both the ubuntu and integration kernels, for example
  559. [18:48:28] * ogra_ac (~ogra@12.157.84.42) Quit (Read error: Connection reset by peer)
  560. [18:48:39] <jayabharath> prpplague: gitiourious tree shoudl be based on integration kernel tree...
  561. [18:48:50] <jayabharath> + patches from ubuntu (picked when make sense)
  562. [18:49:01] <jayabharath> + anything else users want to contribut to a 'non-ubuntu' tree
  563. [18:49:08] <mru> it's not always up to date with the integration tree
  564. [18:49:25] <mru> and please, PLEASE stop favouring ubuntu
  565. [18:50:07] * jayabharath (~jayabhara@nat/ti/x-zfmfaszipmchljeq) Quit (Remote host closed the connection)
  566. [18:50:14] <prpplague> doh
  567. [18:50:16] <armin76> yeah, the sound thing is not funny
  568. [18:50:22] * ogra_ac (~ogra@12.157.84.42) has joined #pandaboard
  569. [18:50:37] <prpplague> armin76: hmm trying to read through it now
  570. [18:50:38] <ogra_ac> grrr
  571. [18:50:41] * ogra_ac curses the LPC network
  572. [18:50:57] <ogra_ac> <ogra_ac> there is a catch22 in that bug i still didnt find a proper solution for yet
  573. [18:50:57] <ogra_ac> <ogra_ac> that makes it a bit hard to fix
  574. [18:50:57] <ogra_ac> <ogra_ac> you can work around that by making sure to first upgrade the kernel to the one from proposed ... then reboot and *only then* upgrade alsa-utils
  575. [18:51:01] * mru curses ogra_ac
  576. [18:51:15] * ogra_ac feels cursed
  577. [18:53:18] <prpplague> armin76: that type of issue is going to be pretty hard for me/us to keep up to date on
  578. [18:54:53] <ogra_ac> i'll fix it properly next week
  579. [18:55:06] <ogra_ac> its a bit hard to work on that while traveling
  580. [18:55:15] <armin76> prpplague: shrug, its just fyi since not everyone is going to use ubuntu
  581. [18:55:29] <ogra_ac> just make sure to install all updates once i uploaded the fix)
  582. [18:55:55] <prpplague> armin76: it is more of a logistics type of thing
  583. [18:55:59] <ogra_ac> for other distros the kernel fix is definitely in the tree
  584. [18:56:17] <ogra_ac> for alsa you have to chheck yourself how your distro is set up
  585. [18:56:36] <armin76> mru: have you built gcc on the panda?
  586. [18:57:05] <ogra_ac> robclark, the pvr driver still needs *lots* of love
  587. [18:57:14] <prpplague> armin76: the gitorious trees were originally intended as validation and used as a place to push patches to other trees such as the ubuntu and omapzoom tress
  588. [18:57:26] <mru> armin76: why would I?
  589. [18:57:47] <ogra_ac> mru, to tell armin76 how it works :)
  590. [18:57:50] <armin76> mru: fun!
  591. [19:00:01] <mru> not my idea of fun
  592. [19:00:10] <robclark> ogra_ac: at least video playback sort of works..
  593. [19:00:36] <ogra_ac> sort of, yeah
  594. [19:00:49] <armin76> mru: well, as always i have problems :P
  595. [19:00:51] <robclark> the resizing issues are actually display driver.. not pvr
  596. [19:00:56] <armin76> mru: i get sigbuses and segfaults
  597. [19:00:56] <ogra_ac> at some point the whole desktop gets so slow that you cant use it though
  598. [19:01:03] <armin76> fun stuff
  599. [19:01:08] <robclark> hmm.. ok.. that is odd..
  600. [19:01:10] <robclark> htop?
  601. [19:01:13] <ogra_ac> i can paint the UI grey by moving a window around
  602. [19:01:24] <robclark> I've not seen that
  603. [19:01:27] <ogra_ac> yeah, X eats some resources
  604. [19:01:35] <ogra_ac> in htop
  605. [19:02:14] <ogra_ac> after i switched to fbdev in xorg.conf it got a bit better
  606. [19:02:22] <ogra_ac> but still slower than plain fbdev
  607. [19:02:47] <ogra_ac> but that way video playback still works at least
  608. [19:02:54] <robclark> hmm.. odd.. I wonder why I haven't seen that..
  609. [19:02:58] <ogra_ac> and the UI is kind of responsible
  610. [19:03:07] <ogra_ac> i havent seen that either before
  611. [19:03:20] <ogra_ac> but i also havent used my E1 excessively yet
  612. [19:03:26] <robclark> ogra_ac: are you using the EFL launcher, or desktop session?
  613. [19:03:32] * robclark uses desktop
  614. [19:03:36] <ogra_ac> and it seems jay has E1's
  615. [19:03:40] <ogra_ac> EFL
  616. [19:03:49] <robclark> hmm.. perhaps that is the difference
  617. [19:03:57] <armin76> E1?
  618. [19:03:59] <ogra_ac> its the presentation machine at the TI booth at LPC
  619. [19:04:03] <FredFred> What' E1?
  620. [19:04:11] <armin76> error1 :D
  621. [19:04:11] <robclark> ogra_ac: use gnome desktop session instead
  622. [19:04:23] <ogra_ac> i'll tell them
  623. [19:04:35] <FredFred> lol...tnx
  624. [19:04:48] <armin76> FredFred: j/k
  625. [19:04:53] <armin76> ogra_ac: you mean A1?
  626. [19:05:20] <ogra_ac> armin76, might be ..... i always forget which letter is striked out :P
  627. [19:05:26] * ssvb (~ssvb@a88-114-220-213.elisa-laajakaista.fi) has joined #pandaboard
  628. [19:06:06] <ogra_ac> but yes, likely A1 ... the final production board
  629. [19:06:46] <J_P> omap4 is capable to capture and encode 1080p right. But that encode is using gstreamer and using the C64x (dsp) right?
  630. [19:07:02] <J_P> don't use the amr processor, just dsp for that right?
  631. [19:07:51] <FredFred> Dumb ?ion, but we're still talking about the pandaboard right?
  632. [19:08:19] <armin76> prpplague: so i understand everyone wanting an updated kernel and uboot+x-loader should look at dev.omapzoom and ignore gitorious, right?
  633. [19:08:31] <armin76> FredFred: we were talking about pandaboard revisions
  634. [19:08:59] <FredFred> ah...kk..thanks
  635. [19:09:35] <armin76> its not about the audi :P
  636. [19:09:58] * jayabharath (~jayabhara@nat/ti/x-jmibcyhcfgsnjqec) has joined #pandaboard
  637. [19:10:19] <ogra_ac> who would want an audi anyway if you can have a panda A1
  638. [19:10:36] <FredFred> yea...you get more HP from a panda! :p
  639. [19:10:48] * ian_brasil_ (~ian_brasi@189.2.128.130) Quit (Ping timeout: 265 seconds)
  640. [19:11:21] <FredFred> Do we even have the audi A1 in the states?
  641. [19:12:32] * cooloney (~roc@12.157.84.6) Quit (Quit: Lost terminal)
  642. [19:14:07] <jayabharath> FredFred: Dont know about A1 - but we have A8's in USA ;)
  643. [19:15:34] <nemesis> and we build them ;p
  644. [19:17:05] <FredFred> Yea...A8 is super sweet if you have a chauffeur! Well..the A8L would be more applicable in that case. Course I can't afford either car or driver.
  645. [19:17:56] <FredFred> No W engine though...*sigh*
  646. [19:18:21] <armin76> cortex a8, uh? :D
  647. [19:18:32] * ogra_ac (~ogra@12.157.84.42) Quit (Ping timeout: 255 seconds)
  648. [19:18:43] * Guest63722 is now known as sum
  649. [19:19:23] <FredFred> referring to audi...srry...and they will have the W12 engine, my mistake
  650. [19:20:09] <wmat> FredFred: what does the A1 cost?
  651. [19:20:20] <wmat> it's not available in Canada either, btw
  652. [19:21:56] * bpadalino (~bpadalino@cpe-72-230-135-225.rochester.res.rr.com) has joined #pandaboard
  653. [19:22:22] * rcn-ee (~voodoo@thief-pool1-30.mncable.net) Quit (Ping timeout: 252 seconds)
  654. [19:23:58] * Monster (~carson@cpe-065-190-191-006.nc.res.rr.com) has joined #pandaboard
  655. [19:24:48] <armin76> FredFred: i was kidding :P
  656. [19:25:11] <FredFred> yea...I can only find it in UK
  657. [19:26:00] * rtos (11ca2aa9@gateway/web/freenode/ip.17.202.42.169) has joined #pandaboard
  658. [19:26:03] <FredFred> UK price starts at 14 (insert pound sign)
  659. [19:26:08] <FredFred> 14k
  660. [19:26:52] <rtos> when is the next rev of pandaboard coming?
  661. [19:27:31] <rtos> or rather is there a timeline for pandaboard revs?
  662. [19:29:01] <armin76> haha
  663. [19:29:11] <armin76> rtos: A1 isn't even shipped yet :P
  664. [19:29:41] <rtos> armin76: yes, I am just worried from my previous experience with beagleboard
  665. [19:30:15] <rtos> armin76: the first couple of revs had an USB issue. I am just wondering if there is any issue with the Si version + board version
  666. [19:30:54] <armin76> rtos: i know the bb history, the ppl were saying so when the bbxm was released :)
  667. [19:31:56] * Monster (~carson@cpe-065-190-191-006.nc.res.rr.com) Quit (Ping timeout: 240 seconds)
  668. [19:33:18] <rtos> from the Pandaboards shipped as of now, are there any known issues?
  669. [19:34:26] <FredFred> Don't know bout issues but Nov 21 is ship date at Digikey...don't see anything about changes.
  670. [19:35:18] <CarsonLynn> have the PEAP boards been shipped yet?
  671. [19:35:32] * armin76 has his from phase1
  672. [19:35:41] <armin76> rtos: i haven't had any issue with mine
  673. [19:35:46] <CarsonLynn> \me is phase 2
  674. [19:35:49] <orbarron> PEAP phase2 should be on the way :D
  675. [19:35:53] <rtos> armin, FredFred: thanks
  676. [19:35:57] <CarsonLynn> too much LaTeX. . .
  677. [19:36:03] <CarsonLynn> orbarron: thanks!
  678. [19:36:12] * rtos (11ca2aa9@gateway/web/freenode/ip.17.202.42.169) has left #pandaboard
  679. [19:36:15] <CarsonLynn> *\LaTeX
  680. [19:39:46] * Monster (~carson@cpe-065-190-191-006.nc.res.rr.com) has joined #pandaboard
  681. [19:42:10] * CarsonLynn (~carson@nom27910d.nomadic.ncsu.edu) Quit (Ping timeout: 252 seconds)
  682. [19:43:12] * ndec (~ndec@nat/ti/x-lxcduhsgdtvovmrk) has joined #pandaboard
  683. [19:43:36] * kapejod (~kapejod@p57BB20C4.dip.t-dialin.net) Quit (Remote host closed the connection)
  684. [19:44:23] * Monster (~carson@cpe-065-190-191-006.nc.res.rr.com) Quit (Ping timeout: 276 seconds)
  685. [19:45:25] * J_P (~neo@189.2.146.53) Quit (Quit: Leaving.)
  686. [19:47:44] * Monster (~carson@cpe-065-190-191-006.nc.res.rr.com) has joined #pandaboard
  687. [19:51:55] * robclark (~robclark@nat/ti/x-ywkstkqbznfdazvn) Quit (Quit: be seeing you)
  688. [19:52:34] * gt11 (~geethat@12.157.84.42) Quit (Quit: Leaving.)
  689. [19:53:06] * rcn-ee (~voodoo@thief-pool1-30.mncable.net) has joined #pandaboard
  690. [19:57:32] * ndec (~ndec@nat/ti/x-lxcduhsgdtvovmrk) Quit (Ping timeout: 255 seconds)
  691. [20:00:23] * RekKumar1 (~a0322918@nat/ti/x-nojssxeyxjsczume) has joined #pandaboard
  692. [20:01:38] * opm (~mike@col06-1-78-231-81-223.fbx.proxad.net) Quit (Quit: Leaving)
  693. [20:03:08] * RekKumar (~a0322918@nat/ti/x-hawpcddcmjaaokwv) Quit (Ping timeout: 240 seconds)
  694. [20:03:54] * robclark (~robclark@nat/ti/x-iduzwovhtzfuziio) has joined #pandaboard
  695. [20:03:54] * ChanServ sets mode +o robclark
  696. [20:05:21] <prpplague> armin76: sorry was in a meeting
  697. [20:06:15] <prpplague> armin76: yes, using either omapzoom integration tree or ubuntu tree will be more actively worked on since there are so many people working on them
  698. [20:06:50] <prpplague> armin76: the primary goal to have 100% of panda support upstream and make the issue irrelevant
  699. [20:09:21] * dk1 (~dkl@unaffiliated/dk1) has joined #pandaboard
  700. [20:10:52] * ndec (~ndec@nat/ti/x-vbzizivbfbwcxzlw) has joined #pandaboard
  701. [20:13:34] * gt11 (~geethat@12.157.84.42) has joined #pandaboard
  702. [20:16:58] * hunt0r (~hunter@dslb-088-064-058-063.pools.arcor-ip.net) has joined #pandaboard
  703. [20:17:43] * gt11 (~geethat@12.157.84.42) Quit (Client Quit)
  704. [20:19:05] * jayabharath (~jayabhara@nat/ti/x-jmibcyhcfgsnjqec) Quit (Remote host closed the connection)
  705. [20:28:55] <zumbi_> hi - is there any panda accesible on-line ?
  706. [20:29:45] <mru> mine is, if you break through my firewall
  707. [20:29:53] <armin76> :D
  708. [20:30:05] <orbarron> zumbi_: not at this time
  709. [20:30:21] <mru> my firewall is set to reject everything from the outside by default
  710. [20:30:46] * NishanthM drives nmap up mru's firewall :P
  711. [20:30:54] <mru> go right ahead
  712. [20:30:57] <NishanthM> :P
  713. [20:31:04] <mru> I doubt you'll find anything
  714. [20:31:10] * NishanthM stopped using nmap since 2002
  715. [20:31:21] <mru> nmap is useful as a diagnostic tool
  716. [20:31:23] * lrg (~lrg@slimlogic.co.uk) Quit (Ping timeout: 245 seconds)
  717. [20:31:38] <mru> or when IT dept decides to move the printer without telling anyone
  718. [20:31:46] <NishanthM> yep... good ol ports :P
  719. [20:34:42] <dk1> orbarron: please post power draw @5v for coremark/coremark2p/some iva test
  720. [20:34:43] * gt11 (~geethat@12.157.84.42) has joined #pandaboard
  721. [20:35:04] <mru> coremark is not a good power benchmark
  722. [20:35:28] <mru> it fits entirely in L1
  723. [20:35:52] <mru> you'll get much higher power draw if you start accessing dram
  724. [20:36:46] <dk1> what do you suggest?
  725. [20:36:55] <mru> what do you want to know?
  726. [20:36:59] <dk1> besides waiting for bdtimark
  727. [20:37:59] <dk1> max draw at various loads, to see how much solar surface area i'd need
  728. [20:38:51] <dk1> faq implies >4A
  729. [20:39:03] <mru> that's probably quite conservative
  730. [20:39:19] <mru> and includes everything on the board
  731. [20:39:22] <dk1> does seem excessive
  732. [20:39:25] <mru> do you need wifi?
  733. [20:39:29] <dk1> yep
  734. [20:39:37] <mru> well, that's going to eat some power
  735. [20:39:53] <mru> probably as much as the omap itself
  736. [20:40:52] <dk1> yes, and all of these numbers should be in the faq :)
  737. [20:41:18] <mru> maybe nobody knows
  738. [20:41:26] <orbarron> will update that in a few..
  739. [20:41:33] <mru> the omap4 datasheet should give some figure for the chip itself
  740. [20:41:42] <mru> then you need to add ram and peripherals
  741. [20:43:15] * orbarron testing minimal-FS with nothing connected then with Ethernet connected then testing with Ubuntu with mouse, keyboard w cpu2 running at 100%
  742. [20:43:35] <mru> 100% of what?
  743. [20:43:41] * jprvita is now known as jprvita|afk
  744. [20:44:12] <mru> a naive busy-loop won't need nearly as much power as a carefully crafted cpuburn loop
  745. [20:44:21] <orbarron> hmm
  746. [20:44:27] * NishanthM (~nmenon@12.157.84.42) Quit (Ping timeout: 250 seconds)
  747. [20:46:18] * NishanthM (~nmenon@12.157.84.42) has joined #pandaboard
  748. [20:50:33] * awayfar (~Adium@nat/ti/x-qaacdykfzflqdwws) Quit (Quit: Leaving.)
  749. [20:51:28] * awayfar (~Adium@nat/ti/x-lyvwvshzdvjkycks) has joined #pandaboard
  750. [20:53:29] * robclark (~robclark@nat/ti/x-iduzwovhtzfuziio) Quit (Remote host closed the connection)
  751. [20:55:35] <dk1> opencv tests would be good
  752. [20:56:10] <mru> dk1: I presume you have a rather specific application in mind when you ask about power
  753. [20:56:43] <mru> if that's the case, you should run that and measure the power consumption
  754. [20:56:48] <mru> then add some margin
  755. [21:02:11] * awayfar (~Adium@nat/ti/x-lyvwvshzdvjkycks) Quit (Quit: Leaving.)
  756. [21:03:16] * ceyusa_ (~vjaquez@155.99.117.91.static.mundo-r.com) has joined #pandaboard
  757. [21:03:37] * ceyusa_ (~vjaquez@155.99.117.91.static.mundo-r.com) Quit (Client Quit)
  758. [21:04:45] * ceyusa (~vjaquez@155.99.117.91.static.mundo-r.com) Quit (Quit: leaving)
  759. [21:05:35] * mru recommends http://article.gmane.org/gmane.comp.lib.glibc.ports/300
  760. [21:05:43] <mru> with a tweak
  761. [21:06:27] <dk1> or repeat what this guy did: http://imxcommunity.org/group/computervisionusingtheimx/forum/topics/optimizing-opencv-face
  762. [21:07:36] <koen> mru: you have that as a non-gmane mangled patch?
  763. [21:07:45] <mru> add /raw to url
  764. [21:07:55] <koen> oooooh
  765. [21:08:00] <mru> beware, it needs a slight tweak to be safe
  766. [21:08:00] <koen> and what's the tweak?
  767. [21:08:44] <mru> __aeabi_memcpy calls normal memcpy
  768. [21:08:58] <mru> but someone said __aeabi_memcpy must preserve _all_ neon regs
  769. [21:09:12] <mru> simplest solution is to save/restore d0-d3 in that memcpy
  770. [21:09:45] <mru> http://hardwarebug.org/files/neon-memcpy.patch
  771. [21:10:53] <koen> you also changed the beq
  772. [21:11:06] <mru> of course I did
  773. [21:11:13] * koen headdesks
  774. [21:11:17] <koen> I see why
  775. [21:11:39] <mru> with that memcpy I get 320MB/s on my panda
  776. [21:11:44] * trem (~trem@mol92-1-81-57-136-23.fbx.proxad.net) has joined #pandaboard
  777. [21:11:44] <mru> compared to 140 with glibc
  778. [21:11:47] <ynezz> hm
  779. [21:11:53] <ynezz> and on beagle?
  780. [21:11:59] <ynezz> classic
  781. [21:12:26] <mru> obviously good code doesn't use memcpy much
  782. [21:12:36] <mru> but good code is rarer than hen's teeth
  783. [21:13:08] * RekKumar1 (~a0322918@nat/ti/x-nojssxeyxjsczume) Quit (Ping timeout: 252 seconds)
  784. [21:13:29] <mru> android memcpy does 277MB/s
  785. [21:18:07] <qq-> mru, good to know you get 320MB/s, but your code will be integrated (somehow) on pandaboards ?
  786. [21:18:28] <mru> that's not my code
  787. [21:18:40] <mru> 285MB/s on beagle btw
  788. [21:18:53] <mru> this is with some panda kernel tweaks too
  789. [21:19:00] <mru> the 320, that is
  790. [21:23:24] <ssvb> mru: on newer cortex-a8 core revisions it seems to be beneficial to write to memory using 64-byte aligned blocks, and on r1pX looked like it did not make much sense to use anything larger than 16-byte alignment
  791. [21:24:19] <mru> the r3 a8 has some changes to the cache system
  792. [21:24:20] <ssvb> mru: so memcpy (or anything memcpy-alike) may need to be tuned even for different revisions of the same core to get best possible performance
  793. [21:24:32] <mru> r1 has exclusive L1 and L2, r3 doesn't
  794. [21:24:45] <mru> a9 is configurable
  795. [21:25:07] <mru> omap4 is non-exclusive for now though
  796. [21:25:14] <mru> it's the damn aux control reg
  797. [21:25:59] <mru> writing full cachelines is probably a good idea if you have write-allocate caches
  798. [21:26:40] <mru> many cpus/caches optimise such stores
  799. [21:27:21] <mru> a9 has a special optimisation for writing full cache lines of zeros
  800. [21:27:26] <mru> also disabled on omap4
  801. [21:27:56] <ssvb> qq-: about your question, I think it's linaro's responsibility to finally push some good memcpy implementation upstream - http://comments.gmane.org/gmane.linux.linaro.toolchain/156
  802. [21:28:33] <qq-> ssvb, cool, thanks to much
  803. [21:29:04] <qq-> 'to get best possible performance' magic worlds ;)
  804. [21:29:46] <mru> marketing words
  805. [21:30:17] <topfs2> mru awesome
  806. [21:30:17] <mru> I suspect few systems ever reach even 50% of their potential
  807. [21:30:53] <topfs2> with the 140 -> 320
  808. [21:31:05] <topfs2> So I guess patching glibc might be valid then?
  809. [21:31:22] <mru> that helps for memcpy-laden code of course
  810. [21:32:07] <ssvb> qq-: the problem with glibc (and libc-ports) is that they require copyright reassignment to FSF, and I just did not want to start these legal procedures without being sure that they don't mind to take a patch from me after that is done (they said that codesourcery had a neon optimized memcpy already)
  811. [21:32:27] <mru> lol x2
  812. [21:32:53] <mru> I wonder how much gnu code is losing due to the copyright crap
  813. [21:35:13] <mru> what's that nonsense people are spouting about neon not being faster on a9?
  814. [21:35:57] <qq-> ssvb, i see ..
  815. [21:36:23] <mru> btw, a simple loop copying 128 bytes at a time with neon gets 435 MB/s
  816. [21:36:53] <ssvb> yes, it's quite possible
  817. [21:37:09] <mru> with some prefetch too
  818. [21:37:31] <ssvb> but what about r1pX cortex-a8?
  819. [21:37:31] <mru> a generic memcpy will of course never be quite as good
  820. [21:38:44] <mru> copying larger blocks seems to be generally faster
  821. [21:39:05] <ssvb> generic memcpy may be as good in some special cases, for example if you are often copying perfectly aligned 4K pages, then it may deserve to become a special case
  822. [21:39:46] <mru> what I mean is that dealing with alignment and odd sizes always costs a little
  823. [21:40:38] <ssvb> yes, sure, that increases prologue/epilogue overhead
  824. [21:40:57] <ssvb> which is bad for small block sizes
  825. [21:41:26] <mru> I think it might be worthwhile to increase the main loop to at least 64 bytes in your code
  826. [21:42:59] * ssvb would happily delegate that to linaro :)
  827. [21:43:09] <mru> good luck with that
  828. [21:43:19] <ds2> how much slower is a DMA based memcpy (just for reference)?
  829. [21:43:27] <mru> linaro is little more than a blame proxy
  830. [21:43:39] <mru> ds2: depends on everything
  831. [21:43:55] <mru> remember that we have a 1MB L2 here...
  832. [21:48:27] <trem> nite all, sweet dreams
  833. [21:48:54] * trem (~trem@mol92-1-81-57-136-23.fbx.proxad.net) Quit (Remote host closed the connection)
  834. [21:51:27] <mru> wow, that linaro one does 425 MB/s
  835. [21:52:39] <mru> it says (c) android ...
  836. [21:52:53] <ssvb> what happens if the relative alignment of source and destination is not perfect?
  837. [21:56:59] * nhg (~a0864305@nat/ti/x-zpyvozennduewwoz) has joined #pandaboard
  838. [21:58:52] * elesueur (~elesueur@120.20.48.219) has joined #pandaboard
  839. [22:00:01] * hunt0r (~hunter@dslb-088-064-058-063.pools.arcor-ip.net) Quit (Remote host closed the connection)
  840. [22:00:17] * dk1 (~dkl@unaffiliated/dk1) has left #pandaboard
  841. [22:03:06] <mru> damn complex cpus
  842. [22:03:12] <mru> everything affects everything
  843. [22:04:06] <av500> linaro memcpy is where?
  844. [22:04:30] <mru> https://launchpad.net/cortex-strings
  845. [22:05:39] <mru> with src and dst x*1M+32k apart, speed goes up
  846. [22:06:16] <av500> mru: thx, yes looks like the android one :)
  847. [22:07:22] <ssvb> mru: maybe some cache aliasing? btw, what is used for indexing data in CPU cache, virtual or physical addresses?
  848. [22:07:34] <mru> L1 is VIPT iirc
  849. [22:07:52] * orbarron (~orb@nat/ti/x-qvhohroywtrzofss) Quit (Quit: Leaving.)
  850. [22:08:00] * prpplague (~danders@nat/ti/x-mkltbwamvaerhalt) Quit (Quit: Leaving)
  851. [22:08:27] <mru> L2 is PIPT
  852. [22:08:50] * ssvb wonders if fragmentation of physical memory affects memcpy performance
  853. [22:08:53] <mru> L2 is 16x 32k ways
  854. [22:09:15] <mru> I doubt fragmentation makes any difference
  855. [22:09:21] <mru> pages are 4k or more
  856. [22:09:39] * gt11 (~geethat@12.157.84.42) Quit (Quit: Leaving.)
  857. [22:11:05] <ssvb> I just wonder if it can create some interesting aliasing patterns
  858. [22:11:13] <mru> so it seems
  859. [22:12:02] <mru> speed is minimum when distance is a multiple of the cache size
  860. [22:13:05] * BThompson (~a0193480@nat/ti/x-rtmcvupggcneyvqa) Quit (Remote host closed the connection)
  861. [22:13:42] <FredFred> http://tastyresearch.files.wordpress.com/2006/10/popvssodamap.png
  862. [22:13:56] <FredFred> srry
  863. [22:15:02] <mru> actually, it seems to dip at multiples of 64k
  864. [22:15:09] <mru> and peaks halfway in between
  865. [22:16:20] <ssvb> mru: I think an interesting experiment could be to artificially cause physical memory fragmentation (maybe by using many mmap/munmap calls in a certain way), and then run memcpy benchmark
  866. [22:17:02] * elesueur (~elesueur@120.20.48.219) Quit (Ping timeout: 255 seconds)
  867. [22:17:23] <mru> easy enough to do by reserving some physmem and mmapping it with MAP_FIXED
  868. [22:17:34] <mru> in some crazy pattern
  869. [22:17:48] <mru> but I'm too lazy to do that just now
  870. [22:17:51] <ds2> are you writing a memory DoS attack? :D
  871. [22:18:15] <mru> the linaro/android memcpy peaks at 490 MB/s
  872. [22:18:41] <mru> which is actually halfway decent
  873. [22:19:07] * ian_brasil_ (~ian_brasi@201-75-18-172-ma.cpe.vivax.com.br) has joined #pandaboard
  874. [22:19:37] <ssvb> ds2: it makes sense to try figuring out some details about how it works, that helps to avoid some of the wtf discoveries later :)
  875. [22:19:57] * symm- (HydraIRC@178.168.29.31) has joined #pandaboard
  876. [22:20:31] <ds2> wtf discoveries are the spice of life
  877. [22:24:06] * nhg (~a0864305@nat/ti/x-zpyvozennduewwoz) has left #pandaboard
  878. [22:26:05] * gxben (~gxben@plo67-1-88-177-29-32.fbx.proxad.net) Quit (Quit: Ex-Chat)
  879. [22:26:53] * NishanthM (~nmenon@12.157.84.42) Quit (Ping timeout: 276 seconds)
  880. [22:30:51] * ian_brasil_ (~ian_brasi@201-75-18-172-ma.cpe.vivax.com.br) Quit (Quit: ian_brasil_)
  881. [22:34:23] * wardred (~wardred@69.109.13.18) has joined #pandaboard
  882. [22:40:16] * mru patches the android/linaro memcpy into his glibc
  883. [22:40:36] <qq-> go go
  884. [22:41:20] * ahmedammar_ (~b33fc0d3@unaffiliated/b33fc0d3) has joined #pandaboard
  885. [22:41:46] <ahmedammar_> topfs2, you around?
  886. [22:45:36] * wardred (~wardred@69.109.13.18) Quit (Remote host closed the connection)
  887. [22:50:38] <topfs2> ahmedammar, yeah
  888. [22:53:28] * elesueur (~elesueur@2402:1800:4000:1:226:b0ff:feea:2790) has joined #pandaboard
  889. [23:02:51] * robclark (~robclark@nat/ti/x-ixlovsqkstqvupya) has joined #pandaboard
  890. [23:02:51] * ChanServ sets mode +o robclark
  891. [23:05:05] * ian_brasil_ (~ian_brasi@201-75-18-172-ma.cpe.vivax.com.br) has joined #pandaboard
  892. [23:19:43] * thaytan (~jan@ppp59-167-167-201.static.internode.on.net) Quit (Quit: Ex-Chat)
  893. [23:20:05] * fandeli (~fandeli@nat/ti/x-nnwavbgucjhrhbjk) Quit (Quit: ChatZilla 0.9.86 [Firefox 3.6.12/20101027124711])
  894. [23:20:34] * ian_brasil___ (~ian_brasi@187.90.112.159) has joined #pandaboard
  895. [23:23:20] * ian_brasil_ (~ian_brasi@201-75-18-172-ma.cpe.vivax.com.br) Quit (Ping timeout: 255 seconds)
  896. [23:35:17] * robclark (~robclark@nat/ti/x-ixlovsqkstqvupya) Quit (Ping timeout: 276 seconds)
  897. [23:51:56] * mpoirier (~quassel@S0106002369de4dac.cg.shawcable.net) Quit (Ping timeout: 240 seconds)

These logs were automatically created by PandaLgBt on irc.freenode.net using the Java IRC LogBot.

See http://microformats.org/wiki/mflogbot for more information.